The challenges and opportunities of artificial intelligence in implementing trustworthy robotics and autonomous systems

摘要

Effective Robots and Autonomous Systems (RAS) must be trustworthy. Trust is essential in designing autonomous and semi-autonomous technologies, because “No trust, no use”. RAS should provide high quality of services, with the four key properties that make it trust, i.e. they must be (i) robust for any health issues, (ii) safe for any matters in their surrounding environments, (iii) secure for any threats from cyber spaces, and (iv) trusted for human-machine interaction. We have thoroughly analysed the challenges in implementing the trustworthy RAS in respects of the four properties, and addressed the power of AI in improving the trustworthiness of RAS. While we put our eyes on the benefits that AI brings to human, we should realise the potential risks that could be caused by AI. The new concept of human-centred AI will be the core in implementing the trustworthy RAS. This review could provide a brief reference for the research on AI for trustworthy RAS.